QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

Vol. 18, Issue 8, 3068-3080, August 2007

This Article Abstract Full Text (PDF) All Versions of this Article: E07-03-0196v1 18/8/3068    most recent Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Lauwers, E. Articles by André, B. Search for Related Content PubMed PubMed Citation Articles by Lauwers, E. Articles by André, B.

Evidence for Coupled Biogenesis of Yeast Gap1 Permease and Sphingolipids: Essential Role in Transport Activity and Normal Control by Ubiquitination

Elsa Lauwers^*, Guido Grossmann, and Bruno André^*

*Laboratoire de Physiologie Moléculaire de la Cellule, Institut de Biologie et de Médecine Moléculaires, Université Libre de Bruxelles, B-6041 Gosselies, Belgium; and University of Regensburg, Cell Biology and Plant Physiology, 93040 Regensburg, Germany

Submitted March 2, 2007; Revised May 18, 2007; Accepted May 29, 2007
Monitoring Editor: Thomas Sommer

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Current models for plasma membrane organization integrate the emerging concepts that membrane proteins tightly associate with surrounding lipids and that biogenesis of surface proteins and lipids may be coupled. We show here that the yeast general amino acid permease Gap1 synthesized in the absence of sphingolipid (SL) biosynthesis is delivered to the cell surface but undergoes rapid and unregulated down-regulation. Furthermore, the permease produced under these conditions but blocked at the cell surface is inactive, soluble in detergent, and more sensitive to proteases. We also show that SL biogenesis is crucial during Gap1 production and secretion but that it is dispensable once Gap1 has reached the plasma membrane. Moreover, the defects displayed by cell surface Gap1 neosynthesized in the absence of SL biosynthesis are not compensated by subsequent restoration of SL production. Finally, we show that down-regulation of Gap1 caused by lack of SL biogenesis involves the ubiquitination of the protein on lysines normally not accessible to ubiquitination and close to the membrane. We propose that coupled biogenesis of Gap1 and SLs would create an SL microenvironment essential to the normal conformation, function, and control of ubiquitination of the permease.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Biological membranes were long seen as a "fluid mosaic" of proteins simply dispersed in a homogeneous lipid bilayer (Singer and Nicholson, 1972). This view contrasts with current models of the plasma membrane that take into account the heterogeneity of lipids in length and structure and the very high density of intrinsic proteins, 30,000 per µm² (Jacobson et al., 2007). These models also integrate the important notion of lateral heterogeneity, i.e., that the plasma membrane contains domains of various lipid compositions and physical properties. Among these domains, lipid rafts have been the subject of many studies. Although the definition of rafts is still debated (Pike, 2006), they are generally viewed as small domains of various sizes (10–200 nm in diameter) that are formed by the dynamic assembly of sphingolipids with sterols (Simons and Ikonen, 1997; Jacobson et al., 2007). These domains selectively include or exclude proteins, and they are thought to exist in a liquid-ordered phase distinct from the bulk liquid-disordered phase of the membrane (Brown and London, 1998). Their tightly packed state makes them more resistant to solubilization by nonionic detergents such as Triton X-100, so that they yield detergent-resistant membranes (DRMs) after detergent extraction in the cold (London and Brown, 2000). Despite the limitations of detergent-dependent methods and the caution that must be exerted when interpreting the resulting data, the isolation of DRMs remains a useful tool for investigating protein–lipid interactions and membrane domains (Brown and London, 1998; Shogomori and Brown, 2003; Lichtenberg et al., 2005). A complementary approach consists in visualizing raft components directly in the plasma membrane of living cells. In Saccharomyces cerevisiae, all the plasma membrane proteins whose detergent solubility has been examined so far have been found in DRMs (Malinska et al., 2004; Lauwers and André, 2006). Yet, proteins recovered in the same DRM can occupy different membrane domains in intact cells (Malinska et al., 2003), raising the debated question of the exact relationship between DRMs in vitro and lipid domains in vivo (London and Brown, 2000; Shogomori and Brown, 2003; Lichtenberg et al., 2005). Tanner and collaborators have indeed shown that two Triton X-100–insoluble proteins, the proton-ATPase Pma1 and the arginine permease Can1, are located in mutually exclusive discrete domains in the plasma membrane of vegetative cells, whereas the hexose permease Hxt1 is homogenously distributed at the cell surface (Malinska et al., 2003). The compartment defined by the presence of Pma1 is called the m embrane compartment occupied by Pma1 (MCP), whereas the membrane compartment occupied by Can1 (MCC) contains several other detergent-insoluble proteins, namely, the uracil permease Fur4 (Malinska et al., 2004), the Sur7 protein (Malinska et al., 2004), the tryptophan permease Tat2 (Grossmann et al., 2007), and the heterologously expressed hexose transporter Hup1 of Chlorella kessleri (Grossmann et al., 2006). Furthermore, plasma membrane structures named eisosomes and proposed to be active sites of endocytosis colocalize with MCC patches (Walther et al., 2006). Both MCC and MCP compartments are stable over time (Malinska et al., 2003), and they are independent of the actin or microtubule cytoskeleton and of the cell wall (Malinska et al., 2004). In a recent study, it was found that the patchy distribution of Hup1, Can1, Fur4, and Tat2 (but not Sur7) is lost upon membrane depolarization (Grossmann et al., 2007). From all these results, it seems that the mechanisms responsible for detergent insolubility and those controlling the lateral distribution of yeast plasma membrane proteins obey to rules that are still poorly understood. According to the lipid shell model (Anderson and Jacobson, 2002), some transmembrane proteins are encased during their biosynthesis into a preassembled lipid microenvironment made of sphingolipids and sterols and corresponding to that of their future host membrane domain. Such lipid shells might confer to these membrane proteins both their detergent insolubility and sorting to specific subdomains of the plasma membrane. Several data sets are consistent with the lipid shell model. For example, a subpopulation of immobilized and disordered lipids is observable by electron-spin resonance in native membranes or lipid–protein systems, but not in protein–free membranes (Lee, 2003). Furthermore, the high-resolution structure of some membrane proteins shows lipid molecules tightly packed or even distorted against the surface of the protein to form a shell (Gonen et al., 2005; Lee, 2005).

The yeast general amino acid permease Gap1 (Grenson et al., 1970) is an attractive system for investigating the mechanisms controlling the membrane trafficking of cell surface proteins. The fate of the Gap1 permease is regulated according to the nitrogen supply conditions of the medium (Magasanik and Kaiser, 2002; Haguenauer-Tsapis and André, 2004). When cells grow on a nitrogen-poor medium, the GAP1 gene is highly expressed (Jauniaux and Grenson, 1990) and the Gap1 protein accumulates at the plasma membrane in a highly active and stable form (Grenson, 1983; De Craene et al., 2001). Endocytosis of Gap1 is triggered when a good nitrogen source such as ammonium or an excess of amino acids is added to the medium (Hein et al., 1995; Stanbrough and Magasanik, 1995; Springael and André, 1998). The presence of these nitrogenous compounds probably inactivates Npr1, a protein kinase essential to protection of Gap1 against ubiquitination (De Craene et al., 2001) and under the control of the TOR signaling pathway (Schmidt et al., 1998). Endocytosis of Gap1 requires prior ubiquitination on two lysine residues present in the N-terminal tail of the permease, at positions 9 and 16 (Springael and André, 1998; Soetens et al., 2001), and closer analysis of this process has revealed that the protein undergoes polyubiquitination with K63-type chains (Springael et al., 1999). The ubiquitination of Gap1 requires the Rsp5 (Npi1) ubiquitin ligase (Hein et al., 1995; Springael and André, 1998) acting together with the Bul1 and Bul2 factors (Yashiroda et al., 1996; Helliwell et al., 2001; Soetens et al., 2001). Delivery of internalized Gap1 into the lumen of the vacuole where the protein is degraded requires its prior sorting into the multivesicular body (MVB) pathway (Nikko et al., 2003), and defects at this level lead either to accumulation of the permease at the late endosome or to effective recycling to the plasma membrane by passage through the Golgi.

In a previous article, we reported that Gap1 fractionates with DRMs when it is present at the plasma membrane (Lauwers and André, 2006), and we proposed that this may be true for all cell surface proteins. Newly synthesized Gap1 acquires detergent insolubility at the Golgi level, and this property is preserved until the protein present at the cell surface is internalized by endocytosis. Moreover, sphingolipid neosynthesis seems essential to the accumulation of newly synthesized Gap1 at the cell surface. When this biosynthesis is defective, neosynthesized Gap1 is targeted to the endosome/vacuolar degradation pathway (Lauwers and André, 2006), as for Pma1 (Bagnat et al., 2001; Gaigg et al., 2005). In the present study, we have further investigated the role of sphingolipids in controlling the intrinsic function, ubiquitination, and membrane trafficking of the yeast Gap1 protein. We have also tested a prediction of the lipid shell model, i.e., that ongoing sphingolipid biosynthesis should be mostly important during Gap1 biogenesis and secretion, and less after Gap1 has integrated the plasma membrane.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Strains, Plasmids, and Media
Unless otherwise indicated, all the S. cerevisiae strains used in this study derive from RH1800 (Table 1). Cells were grown at 24°C in minimal buffered medium (Jacobs et al., 1980) (pH 6.1) with 3% glucose as carbon source, excepted in the confocal microscopy experiment where cells were grown in YNB medium with 3% glucose. To repress GAP1 expression, cells were grown with 0.1% glutamine and 100 mM (NH₄)₂SO₄ as nitrogen sources. A medium containing 0.1% proline as nitrogen source was used to induce Gap1 synthesis. Shifts were performed by filtering cells and resuspending them in medium preheated to the permissive or restrictive temperature. Uracil (0.025%) was added to the medium for uracil-auxotrophic strains. We added D-erythro-dihydrosphingosine (Sigma-Aldrich, Schnelldorf, Germany) solubilized in ethanol to a final concentration of 1 µM, and control cells were treated with ethanol alone. Cell transformation, gene deletions, and excision of the KanMX4 cassette were performed as described previously (Soetens et al., 2001). The plasmids used in this study are listed in Table 1. The YCpJYS12 plasmid was used to complement the his4 and leu2 mutations. Gap1 plasmids were constructed by in vivo recombination between one or two polymerase chain reaction (PCR) fragments carrying either the green fluorescent protein (GFP) or the Gap1–GFP sequence, and, respectively, the YCpGap1 or the YCpFL38 vector. Mutations resulting in the replacement of lysines by arginines were introduced into oligonucleotides used to amplify Gap1 fragments. All plasmids were verified by sequencing. All oligonucleotide sequences are available on request.

Yeast–Cell Extracts and Western Blotting
Total protein extracts were obtained as described previously (Hein et al., 1995). Proteins were resuspended in sample buffer and incubated for 10 min at 37°C before SDS-polyacrylamide gel electrophoresis (PAGE). After transfer to a nitrocellulose membrane (Schleicher & Schüll, Dassel, Germany), the proteins were probed with polyclonal antibodies raised against Gap1 or Pma1 (De Craene et al., 2001). Primary antibodies were detected with horseradish-peroxidase-conjugated anti-rabbit secondary antibody (GE Healthcare, Little Chalfont, Buckinghamshire, United Kingdom) followed by enhanced chemiluminescence (Roche Diagnostics, Mannheim, Germany).

Limited Proteolysis
The experiment was performed essentially as described previously (Gong and Chang, 2001). Cells were lysed with glass beads in 200 µl buffer (50 mM HEPES pH 7.5, NaCl 300 mM). A total membrane fraction was generated by centrifugation at 100,000 x g for 60 min in a SW55 Ti rotor (Beckman Coulter, Fullerton, CA). Membranes were resuspended in buffer and incubated at a protein concentration of 0.5 mg/ml at 30°C and at a trypsin:protein ratio of 1:5. Samples were collected after 0, 0.5, 2, 5, 10, and 20 min, and the reaction was stopped by adding trypsin inhibitor. A control reaction was run for 20 min in the absence of trypsin. Proteins were precipitated by adding 10% trichloroacetic acid (TCA) and Gap1 was analyzed by Western blotting with anti-Gap1 antibodies as described previously. N-tosyl-L-phenylalanine chloromethyl ketone-treated trypsin and trypsin inhibitor type I-S from soybean were purchased from Sigma-Aldrich (Schnelldorf, Germany).

Subcellular Fractionation
Fractionation was performed as described previously (Bagnat et al., 2001; Lauwers and André, 2006). Cells were lysed with glass beads in buffer L (25 mM Tris, pH 8, and 2.5 mM EDTA) supplemented with a cocktail of protease inhibitors (Roche Diagnostics) and 25 mM N-ethylmaleimide (NEM). The cleared lysate was centrifuged at 20,000 x g for 30 min in a SW55 Ti rotor (Beckman Coulter), and the pellet was then resuspended in 20% glycerol in buffer B (10 mM Tris, pH7.4, 0.2 mM EDTA, and 0.2 mM dithiothreitol). The samples (500 µl) were loaded on top of a sucrose step gradient (0.5 ml, 53%; 1 ml, 43%; in buffer B) and centrifuged at 100,000 x g for 2 h in a Beckman SW55 Ti rotor. After centrifugation, six fractions of equal volume were collected from the top of the gradient, and the distribution of Pma1 and Gap1 was analyzed by Western blotting.

Fluorescence Microscopy
Cells transformed with the pEL003 or pEL025 plasmid were laid down on a thin layer of 1% agarose and viewed with a Nikon Eclipse E600 microscope equipped with appropriate fluorescence light filter sets. Images were captured with a Nikon DXM1200 digital camera and processed with Adobe Photoshop CS (Adobe Systems, Mountain View, CA).

Detergent-resistant Membrane Isolation
DRMs were isolated as described previously (Lauwers and André, 2006). Cells were harvested by filtration and lysed in TNE buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, and 5 mM EDTA) supplemented with a cocktail of protease inhibitors (complete protease inhibitors; Roche Diagnostics) and 25 mM NEM. The cleared lysate was incubated with 1% Triton X-100 for 30 min on ice. Lysate (250 µl) was adjusted to 40% iodixanol by adding 500 µl of Optiprep (Axis-Shield, Oslo, Norway), and it was loaded at the bottom of a two-step gradient (1200 µl of 30% iodixanol in TNXE [50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 5 mM EDTA, and 0.1% Triton X-100], and 200 µl TNXE). The gradient was centrifuged at 46,000 rpm for 120 min at 4°C in a Beckman Sw55Ti rotor. Six fractions of equal volume were collected from the top of the gradient, and the proteins were precipitated by incubation with 10% TCA for 30 min. The precipitates were dissolved in 40 µl of 1 M Tris base plus 40 µl of 2X sample buffer (100 mM Tris-HCl, pH 6.8, 4 mM EDTA, 4% SDS, 20% glycerol, and 0.02% bromphenol blue) supplemented with 2% 2-mercaptoethanol. The samples were heated at 37°C for 15 min, and then they were analyzed by Western blotting for the presence of Gap1 and Pma1. Fractions 1–3 were pooled (R, Triton X-100–resistant membranes), as were fractions 4–6 (S, Triton X-100–soluble membranes).

Confocal Microscopy
Confocal sections were scanned using a LSM510-Meta confocal microscope (Zeiss, Jena, Germany). Double-labeled cells (Gap1-green fluorescent protein [GFP] and Sur7-monomeric red fluorescent protein [mRFP]) were sequentially scanned to avoid any cross talk of fluorescence channels. Living cells were immobilized by a thin film (0.5 mm) of 1% agarose and observed.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
The Gap1 Permease Produced in the Absence of Sphingolipid Biosynthesis Is Delivered to the Plasma Membrane but Undergoes Deregulated Ubiquitin-dependent Down-Regulation
We have previously reported that the yeast Gap1 permease, when present at the plasma membrane, fractionates with DRMs enriched in sphingolipids and ergosterol (Lauwers and André, 2006), these being the proposed biochemical equivalent of lipid rafts (Bagnat et al., 2000). Moreover, SL neosynthesis proved essential to accumulation of newly synthesized Gap1 proteins at the cell surface. In a thermosensitive lcb1-100 mutant defective in the first step of SL biosynthesis at the restrictive temperature of 37°C (Zanolari et al., 2000), neosynthesized Gap1 protein, like the Pma1 ATPase (Bagnat et al., 2001), does not accumulate at the cell surface, and it is instead targeted for degradation into the lumen of the vacuole, the lysosome of yeast cells (Lauwers and André, 2006). To determine whether the Gap1 permease synthesized under these conditions reaches the plasma membrane before being degraded, we examined its fate in lcb1-100 cells lacking End3, a protein required for the internalization step of endocytosis (Bénédetti et al., 1994). Cells were first grown at the permissive temperature of 24°C in a medium containing glutamine and ammonium to repress Gap1 synthesis. They were then transferred to proline medium to induce Gap1 production, either at 24°C or at the restrictive temperature of 37°C. Immunoblot analysis showed that Gap1 failed to accumulate to a high level when neosynthesized in lcb1-100 cells shifted to 37°C (Figure 1A), in keeping with previous observations (Lauwers and André, 2006). Yet, Gap1 did accumulate at this temperature in the lcb1-100 end3 double mutant cells (Figure 1A). Marked stabilization of Gap1 was also observed in lcb1-100 cells additionally deficient in Rsp5 (Npi1), the ubiquitin ligase required for ubiquitination and endocytosis of Gap1 (Hein et al., 1995; Springael and André, 1998) (Figure 1A). Because delivery of endocytic cargoes to the vacuolar lumen requires their prior sorting in the MVB pathway (Babst, 2005), we also analyzed the fate of Gap1 neosynthesized in lcb1-100 cells lacking Vps27, a class E Vps factor required for the sorting of cargoes destined for the internal vesicles of MVBs (Bilodeau et al., 2003). Marked stabilization of the permease was also observed in this lcb1-100 vps27 strain (Figure 1A). Hence, when SL biosynthesis is impaired, newly synthesized Gap1 protein seems to be delivered to the plasma membrane before undergoing rapid ubiquitin-dependent down-regulation. These observations were confirmed by examining the localization of a functional Gap1–GFP fusion protein (Figure 1B). When wild-type or end3 cells carrying a plasmid encoding this Gap1–GFP construct were grown in glutamine- and ammonium-containing medium and then transferred to proline-containing medium at either 24 or 37°C for 2 h, fluorescence was observed at the cell surface regardless of the temperature, reflecting accumulation of Gap1–GFP at the plasma membrane. In lcb1-100 cells, in contrast, whereas Gap1-GFP localized correctly to the cell surface when produced at 24°C, only a weak, intracellular fluorescent signal was detected when the permease was induced at 37°C. In the lcb1-100 end3 strain, however, Gap1–GFP neosynthesized at 37°C clearly labeled the cell surface (Figure 1B). This observation corroborates the results of immunoblot analysis (Figure 1A), and it confirms that the permease produced in the absence of SL biosynthesis is delivered to the plasma membrane before being efficiently endocytosed. This conclusion is also supported by the results of a subcellular fractionation experiment, where after a 2-h shift of the lcb1-100 end3 double mutant to proline medium at 37°C, 60% of the Gap1 protein was found with the plasma membrane marker Pma1 in the bottom fractions of a sucrose density gradient (data not shown).

View larger version (50K): [in this window] [in a new window]   Figure 1. Degradation of Gap1 in lcb1-100 mutant cells involves its ubiquitin-dependent endocytosis and MVB sorting. (A) Gap1 synthesis was induced at 24 or 37°C in wild-type (RH1800), lcb1-100 (RH3809), end3 (EL114), end3 lcb1-100 (EL117), npi1 (EL123), npi1 lcb1-100 (EL125), vps27 (EL152) and vps27 lcb1-100 (EL153) cells. Cells were first grown at 24°C in glutamine- and ammonium-containing medium. Aliquots were collected before (t0) and 120 min after the cells were shifted to proline-containing medium, and the presence of Gap1 and Pma1 in total cell extracts was analyzed by Western blotting. The experiments have been carried out at least twice, and proved reproducible; in each case the result of one representative experiment is shown. (B) Wild-type, lcb1-100, end3, and end3 lcb1-100 cells (strains as mentioned above) transformed with YCpGap1-GFP plasmid (pEL003) were observed before (t0) and 120 min after the cells were shifted to proline-containing medium at 24 or 37°C. Left, Nomarski images. Right, GFP fluorescence. Bar, 5 µm. (C) Schematic representation of the sphingolipid biosynthesis pathway in S. cerevisiae. Gene names are shown in italics. DHS-1P, dihydrosphingosine-1-phosphate; PHS-1P, phytosphingosine-1-phosphate; IPC, inositol phosphoceramide; MIPC, mannose inositol phosphoceramide, M(IP)2C, mannose (inositol phosphate)2 ceramide. (D) Gap1 synthesis was induced at 29°C in wild-type (23344c) and lag1 lac1 (EL090) cells. Gap1 activity was measured before and at different times after the beginning of the induction as described in Materials and Methods.

The Gap1 Permease Produced in the Absence of Sphingolipid Neosynthesis Does Not Acquire Detergent Insolubility
The accumulation of Gap1 at the plasma membrane in the lcb1-100 end3 strain enabled us to examine the properties of the permease that is synthesized and delivered to the plasma membrane in the absence of SL neosynthesis. We first examined whether this Gap1 fractionates with DRMs. When Gap1 synthesis was induced in wild-type cells by shifting them to a proline-containing medium at 37°C, the permease was found to fractionate with DRMs (Figure 2A). In contrast, when the permease was similarly induced at 37°C in lcb1-100 end3 double mutant cells, a strong Gap1 signal was detected in the fractions corresponding to detergent-soluble membranes. The residual amount of Gap1 protein induced under the same conditions in the lcb1-100 cells was also found to be soluble in Triton X-100 (Figure 2A). Hence, Gap1 produced in the absence of SL biogenesis, although stabilized at the plasma membrane, fails to fractionate with DRMs. This is surprising, because SLs are very stable in yeast (Dickson and Lester, 2002) and should thus be abundant in the plasma membrane, even if their neosynthesis has been blocked for 2 h, as in the above-mentioned experiment. Accordingly, under the same conditions the proton ATPase Pma1 remained insoluble in Triton X-100 (Figure 2A), indicating that SL-enriched domains were still present at the cell surface but that neosynthesized Gap1 failed to integrate into them. We have previously reported that the ability of Gap1 to associate with DRMs in lcb1-100 cells can be restored by addition of the SL precursor DHS during Gap1 synthesis and secretion (Lauwers and André, 2006). Accordingly, when DHS was present in the growth medium during induction of Gap1 synthesis at 37°C, Gap1 stability and insolubility in Triton X-100 was completely restored in lcb1-100 cells, and the permease also regained full detergent insolubility in lcb1-100 end3 double-mutant cells (Figure 2A).

View larger version (52K): [in this window] [in a new window]   Figure 2. The Gap1 permease synthesized and accumulated at the plasma membrane in the lcb1-100 mutant is inactive and detergent-soluble. (A–C) Wild-type, lcb1-100, end3, and end3 lcb1-100 cells (strains as in Figure 1) were grown at 24°C in glutamine and ammonium medium. Cells were transferred to proline-containing medium at 24 or 37°C, in the presence or the absence of 1 µM DHS. (A) Cells were collected 2 h after the shift at the restrictive temperature and lysed. Gap1 detergent-solubility was examined as described in Materials and Methods. R, Triton X-100–resistant membranes; S, Triton X-100–soluble membranes. (B) Gap1 activity was measured before (t0) and 2 h after the shift to proline-containing medium. The experiments have been carried out at least twice and proved reproducible; in each case, the result of one representative experiment is shown. (C) Gap1–GFP sensitivity to limited proteolysis was examined in end3 and end3 lcb1-100 cells after a 2-h induction at 37°C in the absence or the presence of DHS. Total membranes were prepared and incubated in the absence or presence of trypsin at 30°C for various times, as described in Materials and Methods. Gap1–GFP was analyzed by Western blotting with anti-GFP antibodies. (D) Gap1 sensitivity to limited proteolysis was examined as described above in gap1 and gap1 lcb1-100 cells expressing a nonubiquitinable variant of the permease, Gap19KR, after a 2-h induction at 37°C. Gap19KR was analyzed by Western blotting with anti-Gap1 antibodies.

Down-Regulation of Gap1 Produced in the Absence of Sphingolipid Biosynthesis Involves Its Ubiquitination on Nonclassical Membrane-proximal Lysines
As shown above, the inactive and detergent-soluble Gap1 protein having reached the plasma membrane under conditions preventing SL biosynthesis is rapidly down-regulated in a ubiquitin-dependent manner (Figure 1A). It is noteworthy that this down-regulation occurs although the medium contains only proline as a nitrogen source, i.e., under nitrogen-supply conditions that normally promote stable accumulation of Gap1 at the cell surface, this being dependent on Npr1 kinase (De Craene et al., 2001). Gap1 down-regulation typically occurs when a good nitrogen source such as ammonium is added to the culture medium, conditions proposed to inactivate Npr1 (Schmidt et al., 1998). It thus seems that the control exerted by Npr1 and protecting Gap1 against down-regulation on poor nitrogen media is defective when Gap1 has been produced in the absence of SL biosynthesis.

In wild-type cells, down-regulation of Gap1 caused by the transfer of cells to optimal nitrogen supply conditions involves ubiquitination of the permease on two ubiquitin-acceptor lysines, K9 and K16, located in its amino-terminal tail. A variant permease Gap1^K9,16R, where both lysines have been replaced with arginines, is resistant to ubiquitination and endocytosis, and it is thus protected against degradation (Soetens et al., 2001). We sought to determine whether this Gap1^K9,16R variant is also resistant to degradation in cells defective in SL biosynthesis. Surprisingly, Gap1^K9,16R synthesized in lcb1-100 cells behaved like the native Gap1 protein, i.e., a strong Gap1 signal was immunodetected when the permease was produced at 24°C, but the mutant permease failed to accumulate at the plasma membrane (data not shown) and in immunoblots at the restrictive temperature of 37°C (Figure 3A), reflecting its down-regulation. Because the Rsp5 ubiquitin ligase is nevertheless required for degradation of Gap1 in lcb1-100 cells at 37°C (Figure 1A), we reasoned that other lysines of the permease are probably ubiquitinated. We thus constructed a mutant permease, Gap1^9KR, with all nine lysine residues of the N-terminal tail replaced with arginines. When produced in wild-type cells at 24 or 37°C, the Gap1^9KR protein was normally secreted to the plasma membrane, and it was active at both temperatures (Figure 3, B–D). In the lcb1-100 mutant at the restrictive temperature of 37°C, in contrast to the native permease and the Gap1^K9,16R variant, the newly synthesized Gap1^9KR protein seemed stable (Figure 3E). Furthermore, Gap1^9KR produced in wild-type or lcb1-100 cells was found at the plasma membrane regardless of the temperature, as indicated by visualization of a functional Gap1^9KR–GFP fusion protein at the cell surface (Figure 3B) and by cofractionation of Gap1^9KR with the plasma membrane marker Pma1 in a sucrose density gradient (Figure 3C). Yet the cell surface-located Gap1^9KR proved inactive in the lcb1-100 mutant at 37°C (Figure 3D), like the native Gap1 permease trapped at the plasma membrane in the lcb1-100 end3 mutant (Figure 2B). This again argues for the association of Gap1 with SL-enriched domains being crucial to the transport activity of the permease.

View larger version (52K): [in this window] [in a new window]   Figure 3. Degradation of Gap1 permease in the lcb1-100 mutant involves its ubiquitination on additional membrane-proximal lysine residues. Cells of the gap1 (EL041) and gap1 lcb1-100 (EL103) strains were grown at 24°C in glutamine and ammonium-containing medium, and transferred at 24 or 37°C in medium containing proline as the sole nitrogen source to induce Gap1 expression. (A) Cells expressing plasmid-borne Gap1 or Gap1K9,16R. Aliquots were collected before (t0) and 120 min after the shift, and the presence of Gap1 and Pma1 in total cell extracts was analyzed by Western blotting. (B) Cells expressing plasmid-borne Gap1–GFP or Gap19KR–GFP were observed before (t0) and 120 min after the shift to proline-containing medium. Left, Nomarski images. Right, GFP fluorescence. Bar, 5 µm. (C) Cells expressing plasmid-borne Gap1 or Gap19KR were lysed and fractionated on a sucrose step gradient as described in Materials and Methods. The distribution of Pma1 and Gap1 was analyzed by Western blotting. Fractions 5 and 6 correspond to the plasma membrane, whereas fractions 2–4 correspond to the ER, Golgi, and vacuolar membranes. The intracellular pool of Gap1 is expected to represent newly synthesized proteins that are present in the secretory pathway. (D) Permease activity was quantified in cells expressing plasmid-borne Gap1 or Gap19KR, before (t0) and 120 min after the shift to proline medium. Activity measurements were carried out at least twice, and they proved reproducible; the result of one representative experiment is shown. (E) Cells expressing various plasmid-borne Gap1–GFP mutant forms. Aliquots were collected 120 min after the shift to proline-containing medium, and the presence of Gap1 and Pma1 in total cell extracts was analyzed by western blotting. (F) Schematic representation of the Gap1 N-terminal region. The first transmembrane segment of the permease (predicted by TMHMM; CBS, Lyngby, Denmark) is represented by a gray cylinder. Lysines are indicated in bold, and lysines that are targets for ubiquitination in lcb1-100 cells are enlarged.

Sphingolipid Neosynthesis Is Specifically Required during Gap1 Production
The above-mentioned data show that arrest of SL neosynthesis during Gap1 production and secretion results in lack of Gap1 activity, increased sensitivity of the permease to trypsinolysis, inability of Gap1 to integrate existing SL-enriched domains (containing other proteins such as Pma1), and unregulated Gap1 down-regulation. We next wondered what would be the effect of an arrest of SL neosynthesis occurring after Gap1 protein has accumulated at the plasma membrane. We started with lcb1-100 cells expressing Gap1^9KR–GFP and grown at the permissive temperature of 24°C with proline as the sole nitrogen source. Under these conditions, the permease accumulated at the plasma membrane in an active (data not shown) and stable form, and it fractionated largely with DRMs (Figure 4A), as expected. A significant amount of Gap1^9KR was also detected in the fractions corresponding to detergent-soluble membranes, probably corresponding to permeases en route to the plasma membrane and having not yet integrated SL-enriched domains. Accordingly, this soluble pool of Gap1^9KR was completely resorbed after permease synthesis was arrested and the cells grown for two more hours (Figure 4A). Gap1^9KR–GFP production was then blocked, and SL neosynthesis was arrested by transferring the cells to glutamine- and ammonium-containing medium at 37°C. No major change in permease detergent-solubility or activity was observed even 2 h after the shift to the restrictive temperature (Figure 4A; data not shown). Thus, once Gap1 has established its location within SL-enriched domains at the plasma membrane, there is no need for sustained SL neosynthesis to maintain this association, at least during the first hours after Gap1 production. This resembles the behavior of the proton ATPase Pma1, which remains detergent-insoluble even 2 h after arrest of SL neosynthesis (e.g., Figure 2A). We next performed the reverse experiment, preaccumulating Gap1^9KR at the plasma membrane outside SL-enriched domains and then allowing SL neosynthesis to resume. For this, lcb1-100 mutant cells grown at 24°C were first shifted to proline-containing medium at 37°C to induce synthesis of Gap1^9KR. As expected, Gap1^9KR accumulated at the plasma membrane but was inactive (data not shown) and totally soluble in Triton X-100 (Figure 4B). Gap1^9KR synthesis was then blocked by transferring the cells to glutamine- and ammonium-containing medium, and 30 min later the SL precursor DHS was added to the culture to restore SL biosynthesis. Remarkably, the permease remained excluded from DRMs (Figure 4B) and inactive (data not shown), even 2 h after DHS addition, indicating that a restored SL supply cannot reinstate either the activity of the permease or its association with SL-enriched domains. Together, these results suggest that SL biosynthesis is crucial specifically during synthesis and secretion of the permease. Accordingly, maintenance of this biosynthesis is not required to preserve the properties of Gap1 having reached the plasma membrane, and the defects displayed by Gap1 produced in the absence of SL biosynthesis seem irreversible, i.e., they are not compensated by restoring normal SL biosynthesis. This suggests that the biogenesis of Gap1 is somehow coupled to that of SLs, this being crucial for Gap1 to acquire detergent insolubility, activity, and normal control of its ubiquitination.

View larger version (65K): [in this window] [in a new window]   Figure 4. Sphingolipid neosynthesis is required during Gap1 production and secretion, but not once the permease has reached the plasma membrane. (A) gap1 and gap1 lcb1-100 cells expressing plasmid-borne Gap19KR–GFP were grown at 24°C in proline medium and then transferred for a 30-min incubation in glutamine- and ammonium-containing medium at 24°C to repress Gap1 synthesis. Next, cells were transferred to glutamine- and ammonium-containing medium at 37°C. Gap1 detergent solubility was examined in proline-grown cells and 2 h after the shift to the restrictive temperature, as described in Materials and Methods. R, Triton X-100–resistant membranes; S, Triton X-100–soluble membranes. (B) gap1 and gap1 lcb1-100 cells expressing plasmid-borne Gap19KR were grown at 24°C in glutamine and ammonium medium, and then they were transferred to proline medium at 37°C for 2 h. Gap1 synthesis was then repressed by shifting back to glutamine and ammonium medium, and DHS was added to the culture 30 min later. Gap1 detergent-solubility was examined 2 h after induction of permease synthesis and 1 or 2 h after addition of DHS, as described in Materials and Methods.

View larger version (74K): [in this window] [in a new window]   Figure 5. Gap1–GFP is homogenously distributed at the plasma membrane. Localization of the Gap1 permease by confocal microscopy. (A) Cross sections (top row) and surface (middle row) views of GYS180 cells expressing Gap1–GFP (green in merge) and the MCC marker Sur7–mRFP (red in merge). Cells were grown on proline-containing medium as a steady-state condition. The intensity profile along the dashed arrow (merge image) shows no significant correlation between the fluorescence signal of Gap1–GFP and the distinct intensity peaks of Sur7–mRFP. Bar, 5 µm. (B) Cross sections of cells expressing Gap1–GFP visualized after a 2-h induction in proline-containing medium. Wild-type (RH1800) cells were induced at 24 or 37°C, whereas end3 (EL114) and end3 lcb1-100 (EL117) cells were induced at 37°C. For better comparability of the protein distribution at the plasma membrane the images were adjusted to approximately the same total intensity. Bar, 5 µm.

The homogeneous distribution of Gap1 at the cell surface is reminiscent of the situation described for the hexose permease Hxt1 (Malinska et al., 2003). The Gap1 and Hxt1 permeases thus seem to define a third group of yeast plasma membrane proteins, i.e., those that do not specifically segregate into the MCC or MCP. Furthermore, Gap1 plasma membrane repartition is not altered if the permease is produced in the absence of SL biosynthesis.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
The main conclusion of this work is that de novo SL biosynthesis is crucial during Gap1 synthesis and/or secretion for the permease to be functional and stable at the plasma membrane. The permease synthesized in the lcb1-100 mutant at the restrictive temperature of 37°C, i.e., in the absence of SL neosynthesis, is delivered to the cell surface, but it is rapidly down-regulated. Using mutations stabilizing the permease at the cell surface, we have shown that Gap1 synthesized under these conditions displays several abnormalities, including a complete lack of activity, an increased sensitivity to proteolysis, and an inability to fractionate with detergent-resistant membranes. These defects can be prevented if SL biogenesis is restored in lcb1-100 cells by providing the SL precursor DHS during Gap1 synthesis and secretion, but they cannot be compensated by adding DHS after Gap1 has reached the plasma membrane. Conversely, once Gap1 has integrated the plasma membrane, there is no need for sustained SL synthesis to preserve its functional properties, at least during the first 2 h. We propose that these observations are most easily interpretable in terms of a model where the Gap1 permease associates with neosynthesized SLs during its production and/or secretion, this association shaping the permease into a conformation essential to activity. Lack of this association would irreversibly distort neosynthesized Gap1 into a conformation incompatible with its function. This permease would also be unable to associate with the abundant and stable SLs present at the plasma membrane, or to integrate SL-enriched domains, this probably accounting for its inability to fractionate with DRMs. That intrinsic plasma membrane proteins tightly associate with lipids is at the heart of the "lipid shell" model proposed by Anderson and Jacobson (2002). This model suggests that some transmembrane proteins have an affinity for sterols and SLs and that they become surrounded by a shell of these lipids during their production. Because of the molecular compatibility between lipids of the shell and those present in rafts, these proteins would be inclined to associate with pre-existing rafts. This might explain why lipid rafts selectively include (or exclude) some proteins, thus acquiring specific functional properties (Anderson and Jacobson, 2002). Formation of an SL shell has also been proposed to strongly influence the conformation of the encased protein (Fantini, 2003). More generally, the lipid composition of the bilayer is thought to be a determinant of the activity and structure of intrinsic membrane proteins (Lee, 2004; Opekarova et al., 2005; Zhang et al., 2005). Interestingly, lipids forming an annular shell around transmembrane proteins have been revealed in some high-resolution structural studies, e.g., in the case of the lens-specific aquaporin-0 (Gonen et al., 2005). In this case, tight packing of lipid fatty acyl chains against the rough surface of the protein was apparent, some of these chains being distorted to correctly wrap around the bulky side chain of the protein. This tight association probably contributes to the permeability barrier properties of the lipid bilayer in direct contact with the protein surface (Lee, 2005).

As mentioned above, the proposed association of proteins with a lipid shell might also account for their lateral distribution into domains of the plasma membrane (Anderson and Jacobson, 2002). In yeast, multiple permeases (Can1, Fur4, Tat2) and other membrane proteins (e.g., Sur7) colocalize in discrete patches constituting the MCC, whereas Pma1 is excluded from this compartment and apparently occupies the remaining space of the cell surface, called the MCP (Malinska et al., 2003; Malinska et al., 2004; Grossmann et al., 2006). We have shown in this work that the Gap1 permease is instead homogenously distributed at the plasma membrane and thus displays a lateral segregation profile very similar to that of the Hxt1 hexose transporter (Malinska et al., 2003). Because all the above-cited proteins fractionate with DRMs (Dupré and Haguenauer-Tsapis, 2003; Malinska et al., 2003, 2004; Umebayashi and Nakano, 2003; Lauwers and André, 2006), the question arises of why they segregate differently in the plasma membrane. One possibility is that all these proteins are detergent-insoluble because they are encased in a lipid shell during their biosynthesis and secretion. The lipid composition and structure of these shells, however, could differ substantially according to the protein, conferring specific aggregation properties that would determine protein self-distribution between distinct membrane domains (Figure 6). In support of this view, it is noteworthy that the protein components of the MCP and MCC show distinct lipid requirements. For example, biosynthesis of SLs but not ergosterol is essential to cell surface accumulation of Pma1 (Gaigg et al., 2005), whereas ergosterol, which is more abundant in the MCC than in the MCP (Grossmann et al., 2007), is necessary for plasma membrane targeting of Tat2 (Umebayashi and Nakano, 2003). It has also recently been reported that the patchy distribution of some proteins normally present in the MCC is lost upon membrane depolarization (Grossmann et al., 2007). This depolarization might affect the conformation of the intrinsic membrane proteins (Patzlaff et al., 2002) and consequently their association with surrounding lipids and their ability to aggregate preferentially with other proteins.

View larger version (25K): [in this window] [in a new window]   Figure 6. A model for lateral distribution of yeast plasma membrane proteins. Left, confocal surface view a yeast cell expressing Sur7–GFP to reveal the MCC. Middle, schematic representation of a single MCC patch surrounded by MCP. Right, hypothetical representation of proteins distributed in the MCC and/or MCP. Proteins (white circles) are encased in a lipid shell (ring around white circle) of various composition and structure and that determines the segregation properties of proteins. Some proteins self-aggregate via their lipid shell to form MCC patches (proteins with white rings). These MCC proteins cannot aggregate with Pma1 nor (possibly) with other proteins (marked with black rings) that self-assemble to form the MCP. Other proteins such as Gap1 and Hxt1 (marked with gray rings) are surrounded by a lipid shell compatible with those encasing both MCC and MCP proteins, and thus they do not specifically segregate in MCC or MCP.

Our data also suggest an important role of SL biogenesis during Gap1 production and secretion for proper control of the permease by ubiquitination. Under normal conditions, Gap1 undergoes ubiquitination on N-terminal lysines at positions 9 and 16 only once the control exerted by the Npr1 kinase has been released, i.e., when a favored nitrogen is added to the medium (De Craene et al., 2001; Soetens et al., 2001). In contrast, Gap1 proteins produced in the absence of SL neosynthesis escape this Npr1-dependent control and undergo efficient Rsp5-mediated down-regulation even in cells growing on a poor nitrogen source. Although this down-regulation can rely solely on the classical ubiquitin acceptor lysines 9 and 16 in the N-terminal tail, our data show that three additional lysines of this tail, at positions 76, 87, and 91, can also promote the unregulated down-regulation of Gap1. Under normal conditions of SL biogenesis, these three residues, located close to the membrane (Figure 3F), have so far not been implicated in nitrogen-regulated control of Gap1 ubiquitination. Yet in two global analyses of ubiquitin conjugates in yeast, Gap1 was found to be ubiquitinated at lysine 76 (Hitchcock et al., 2003; Peng et al., 2003). In both studies, yeast cells were grown on a standard rich medium, conditions under which GAP1 gene transcription is largely repressed and Gap1 permease synthesized to a limited level is directly sorted from the Golgi to the endosome/vacuolar degradation pathway without passing through the plasma membrane (Magasanik and Kaiser, 2002). It is thus possible that ubiquitination of Gap1 at position 76 might also occur under particular environmental conditions, even when SL biogenesis occurs normally. In one of the above-cited proteomics studies, the analyzed cells contained a mutation impairing degradation of protein substrates of the ER-associated degradation pathway (Hitchcock et al., 2003), raising the possibility of a role of lysine 76 (and possibly lysines 87 and 91) in ubiquitination of Gap1 at ER level. Whatever the membrane compartment where ubiquitination of neosynthesized Gap1 takes place when SL biogenesis is defective, our results suggest that this modification is involved in a quality-control mechanism that ensures removal of misfolded permeases from the plasma membrane. Accordingly, Gap1 produced in the absence of SL biosynthesis is more susceptible to trypsinolysis, suggesting that the permease conformation is altered (Figure 2, C and D). SLs might thus act as chaperones to shape newly synthesized Gap1 into its correct conformation, similarly to the role of phosphatidylethanolamine in proper folding of the -aminobutyric acid permease GabP of Escherichia coli (Zhang et al., 2005). Alternatively, a more direct role of SLs in protecting Gap1 against ubiquitination could also be considered. Although mammalian SLs are known to reside mainly in the exoplasmic leaflet of the plasma membrane (Holthuis et al., 2001), 20% of them are found in the cytoplasmic leaflet (Boon and Smith, 2002). Even if no data are currently available about the distribution of SLs over the two leaflets of the yeast plasma membrane (van Meer and Holthuis, 2000), it is thus conceivable that at least a small fraction of these lipids face the cytoplasm, in particular if they are forming a shell around transmembrane proteins. Because the cytosolic lysines 76, 87, and 91 of Gap1 are close to the membrane (Figure 3F), they might be part of the region containing the SL head groups and thus be inaccessible to posttranslational modifications. When Gap1 is not associated with SLs, these lysines would in contrast be exposed and targeted by the Rsp5 ubiquitin ligase, promoting degradation of the permease. Although this unregulated process might be considered a fortuitous consequence of impaired SL biogenesis, it might also be regarded as an actual quality-control mechanism ensuring that only Gap1 proteins correctly associated with SLs (and thus in an active conformation) can be stabilized at the cell surface.

SL biogenesis is also important for plasma membrane accumulation of several other proteins, including the Pma1 H⁺ ATPase (Bagnat et al., 2001; Eisenkolb et al., 2002; Wang and Chang, 2002; Gaigg et al., 2005). In this case, a role of SLs in protein oligomerization has also been proposed (Wang and Chang, 2002). Furthermore, a mutant form of Pma1 (Pma1-10) that fails to associate with DRMs is normally targeted to the cell surface, but it undergoes rapid and ubiquitin-dependent down-regulation (Wang and Chang, 2002; Liu and Chang, 2006). Cell surface delivery of several other plasma membrane proteins is also defective in lcb1-100 cells (Sutterlin et al., 1997; Hearn et al., 2003; Dupré and Haguenauer-Tsapis, 2003), although sphingoid base synthesis is not a general requirement for trafficking along the secretory pathway (Sutterlin et al., 1997). In the Tat2 tryptophan permease, ergosterol is essential to targeting and accumulation at the plasma membrane (Umebayashi and Nakano, 2003). The picture emerging from these studies and the present work points to a crucial role of the lipid microenvironment of plasma membrane proteins in determining their intrinsic conformation and function, controlling their ubiquitination, and probably in determining their lateral distribution between distinct membrane domains.

	ACKNOWLEDGMENTS

 
We are grateful to H. Riezman (University of Geneva, Switzerland) for providing strains, to Widmar Tanner for critical comments on the manuscript and for support, and to Rosine Haguenauer-Tsapis and Jean-Marie Ruysschaert for careful reading of the manuscript. We thank all the members of the laboratory for helpful discussion in the course of this work. This work was supported by grant FRSM 3.4.605.0 [EC] 5. F from the National Fund for Scientific Research, Belgium; the Belgian Fortron Fund; and the Communauté Française de Belgique (Action de Recherche Concertée grant 04/09-307). E.L. is a Research Fellow of the National Fund for Scientific Research, Belgium.

	Footnotes

 
This was published online ahead of print in MBC in Press (http://www.molbiolcell.org/cgi/doi/10.1091/mbc.E07-03-0196) on June 6, 2007.

Address correspondence to: Bruno André (bran{at}ulb.ac.be).

Abbreviations used: DHS, dihydrosphingosine; DRM, detergent-resistant membrane; PHS, phytosphingosine; SL, sphingolipid.

	REFERENCES

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Anderson, R.G.W., and Jacobson, K. (2002). A role for lipid shells in targeting proteins to caveolae, rafts, and other lipid domains. Science 296, 1821–1825.[Abstract/Free Full Text]

Babst, M. (2005). A protein's final ESCRT. Traffic 6, 2–9.[CrossRef][Medline]

Bagnat, M., Chang, A., and Simons, K. (2001). Plasma membrane proton ATPase Pma1p requires raft association for surface delivery in yeast. Mol. Biol. Cell 12, 4129–4138.[Abstract/Free Full Text]

Bagnat, M., Keränen, S., Shevchenko, A., Shevchenko, A., and Simons, K. (2000). Lipid rafts function in biosynthetic delivery of proteins to the cell surface in yeast. Proc. Natl. Acad. Sci. USA 97, 3254–3259.[Abstract/Free Full Text]

Bénédetti, H., Raths, S., Crausaz, F., and Riezman, H. (1994). The END3 gene encodes a protein that is required for the internalization step of endocytosis and for actin cytoskeleton organization in yeast. Mol. Biol. Cell 5, 1023–1037.[Abstract]

Bilodeau, P. S., Winistorfer, S. C., Kearney, W. R., Robertson, A. D., and Piper, R. C. (2003). Vps27-Hse1 and ESCRT-I complexes cooperate to increase efficiency of sorting ubiquitinated proteins at the endosome. J. Cell Biol 163, 237.[Abstract/Free Full Text]

Bonneaud, N., Ozier-Kalogeropoulos, O., Li, G. Y., Labouesse, M., Minvielle-Sebastia, L., and Lacroute, F. (1991). A family of low and high copy replicative, integrative and single-stranded S. cerevisiae/E. coli shuttle vectors. Yeast 7, 609–615.[CrossRef][Medline]

Boon, J. M., and Smith, B. D. (2002). Chemical control of phospholipid distribution across bilayer membranes. Med. Res. Rev 22, 251–281.[CrossRef][Medline]

Brown, D. A., and London, E. (1998). Structure and origin of ordered lipid domains in biological membranes. J. Membr. Biol 164, 103–114.[CrossRef][Medline]

De Craene, J.-O., Soetens, O., and André, B. (2001). The Npr1 kinase controls biosynthetic and endocytic sorting of the yeast Gap1 permease. J. Biol. Chem 276, 43939–43948.[Abstract/Free Full Text]

Dickson, R. C., and Lester, R. L. (2002). Sphingolipid functions in Saccharomyces cerevisiae. Biochim. Biophys. Acta 1583, 13–25.[Medline]

Dupré, S., and Haguenauer-Tsapis, R. (2003). Raft partitioning of the yeast uracil permease during trafficking along the endocytic pathway. Traffic 4, 83–96.[CrossRef][Medline]

Eisenkolb, M., Zenzmaier, C., Leitner, E., and Schneiter, R. (2002). A specific structural requirement for ergosterol in long-chain fatty acid synthesis mutants important for maintaining raft domains in yeast. Mol. Biol. Cell 13, 4414–4428.[Abstract/Free Full Text]

Fantini, J. (2003). How sphingolipids bind and shape proteins: molecular basis of lipid-protein interactions in lipid shells, rafts and related biomembrane domains. CMLS 60, 1027–1032.[Medline]

Futerman, A. H., and Riezman, H. (2005). The ins and outs of sphingolipid synthesis. Trends Cell Biol 15, 312–318.[CrossRef][Medline]

Gaigg, B., Timischl, B., Corbino, L., and Schneiter, R. (2005). Synthesis of sphingolipids with very long chain fatty acids but not ergosterol is required for routing of newly synthesized plasma membrane ATPase to the cell surface of yeast. J. Biol. Chem 280, 22515–22522.